In the existing self-excitation push-pull type converters, their circuit structure is based on the DC converter of self-excitation push-pull transistor single transformer invented by G. H. Royer in the United States in 1955, it is also referred to as Royer circuit, which is the start to realize high frequency conversion control circuit; some circuits are based on the self-excitation push-pull dual transformer circuit of Jensen of the United States in 1957, which was later known as self-oscillating Jensen circuit or Jensen circuit; both of these circuits were later referred to as self-excitation push-pull converter. The working principle of the self-excitation push-pull converter is described on pp. 67˜70 of Principle and design of switching power source published by the Electronic Industry Press, the ISBN No. of the book is 7-121-00211-6. The circuits are mainly in the forms of the above-mentioned famous Royer circuit and self-oscillating Jensen circuit.
Shown in FIG. 1-1 is a common application of the self-excitation push-pull converter, it is based on Royer circuit; shown in FIG. 1-2 is the famous self-oscillating Jensen circuit; in FIG. 1-1 and FIG. 1-2, the circuits oscillate by using the magnetic core saturation characteristics of transformer B1, in the Jensen circuit of FIG. 1-2, the self-oscillating frequency and drive function of the circuit is realized by the magnetic saturated transformer B1, therefore, the main power transformer B2 can work in a non-saturated state.
The oscillation frequency of the Royer circuit is a function of the power source voltage, which is described in Line 18 on p. 68 of Principle and design of switching power source published by the Electronic Industry Press, the ISBN No. of the book is 7-121-00211-6. It is quoted as follows:
                    f        =                              Vs                          4              ⁢                                                          ⁢              BwSN                                ×                      10            4                    ⁢                                          ⁢          Hz                                    Formula        ⁢                                  ⁢                  (          1          )                    where: f is the oscillation frequency; Bw the working magnetic induction intensity (T), normally taken as 50%˜70% of the magnetic saturation point Bm value; N is the No. of coil turns; S the magnetic core effective sectional area; and VS the working power source voltage.
To better understand the working principle of the Royer circuit, especially the oscillation with magnetic core saturation characteristics, its working principle is described with FIG. 1-1 as an example.
The circuit in FIG. 1-1 is in such a structure: the input filtering capacity C is connected between the voltage input and ground, to filter the input voltage; the input voltage after filtration is connected to the start circuit, which is formed by the shunted biasing resistor R1 and capacitor C1, the two ends of the biasing resistor R1 are respectively connected with the voltage input and the central taps of the primary side coils NB1 and NB2 of transformer B1 which provides positive feedback to the bases of the two push-pull transistor TR1 and TR2, the emitters of the two push-pull transistor TR1 and TR2 share a ground, the two collectors are respectively connected to the two ends of the transformer primary side coils NP1 and NP2, the bases are connected to the two ends of the transformer primary side coils NB1 and NB2, and the central tap in the primary side coils NP1 and NP2 are connected to the voltage input; the secondary side coil NS of transformer B1 connects the output circuit to the voltage output.
The working principle can be briefly described as: refer to FIG. 1-1, at the moment when the power is turned on, the shunted circuit of biasing resistor R1 and capacitor C1 provides a forward bias for the base and emitter of the transistors TR1 and TR2 via windings NB1 and NB2, the two transistors TR1 and TR2 start to conduct. As the characteristics of the two transistor cannot be completely the same, one of them will become conducting first, suppose transistor TR2 becomes conducting first, and produces the collector current IC2, the voltage in the corresponding NP2 winding is positive at the top and negative at the bottom, according to the dotted terminal relationship, an induced voltage positive at the top and negative at the bottom will also appear at its base coil NB2, this voltage increases the base current of transistor TR2, which is a forward feedback process, therefore it quickly makes transistor TR2 saturated and conducting. Similarly, the voltage of the coil NB1 corresponding to transistor TR1 is positive at the top and negative at the bottom, and it reduces the base current of transistor TR1, so that this transistor is soon completely cut off.
The current in coil NP2 winding corresponding to transistor TR2 and the magnetic induction intensity produced by this current increases linearly with time, but when the magnetic induction intensity increases to approach or reach the saturation point Bm of the transformer B1 magnetic core, the induction in the NP2 will decrease quickly, resulting in sharp increase of the collector current of transistor TR2, at a rate much higher than the increasing rate of base current, the transistor TR2 becomes unsaturated, the voltage drop Uce across the collector and emitter of transistor TR2 increases. Correspondingly, the voltage on transformer NP2 winding reduces by the same value, and the voltage induced in coil NB2 winding reduces, resulting in reduction of the base voltage of transistor TR2, so that transistor TR2 changes in the direction of cut-off. At this moment, the voltage in the coil of transformer B1 will reverse, to make the other transistor TR1 conduct, and after that, this process is repeated, to form push-pull oscillation. The waveform at the winding Ns output end is as shown in FIG. 2-1, it can be seen that except the “collector resonance Royer circuit” that outputs sinusoidal wave or approximate sinusoidal wave, the working waveform of the self-excitation push-pull converter is close to a square wave. The collector resonance Royer circuit is also called “cold cathode fluorescent lamp inverter”, referred to as CCFL inverter or CCFL converter, the CCFL converter is connected in series an inductor with a inductance over ten times that of the main power winding in the power supply circuit, to obtain an output of sinusoidal wave or approximate sinusoidal wave. FIG. 3 is the square hysteresis line of the magnetic core of transformer B1, on which +Bm and −Bm are the two magnetic saturation points of the magnetic core. In a half cycle as shown in FIG. 2-1, the moving line of the working point of the magnetic core of transformer B1 is ABCDE, and its moving line in the next half cycle is EFGHA. In fact, when the current in coil windings corresponding to transistors TR2 or TR1, and the magnetic induction intensity produced by this current increase linearly with time to point D or H as shown in FIG. 3, the circuit will perform push-pull conversion, i.e. when one transistor conducts, the opposite transistor cuts off. As transistors have a storage time, that is, after the transistor base has received a cutting off signal, the collector current will start to drop with a short time delay till cut-off, the storage time can occur in FIG. 3, the moving line of the magnetic core working point will be from point D to E, correspondingly, the moving line of the magnetic core working point will move from point H to A.
Its feature is: push-pull oscillation is conducted by using the saturation characteristics of the magnetic core, the transformer output waveform is approximate square wave, and the circuit conversion efficiency is fairly high. As the magnetic core should become nearly saturated at the specific time moment, a magnetic core with air gap cannot be used. A self-excitation push-pull converter must use a magnetic saturated magnetic core, and magnetic core plus air gap is a generally known means to resist magnetic saturation.
In this literature, magnetic core, as in other generally known literatures, refers to a ferrite material, i.e. a sintered magnetic metal oxide of the mixture of various ferrite oxides, and magnetic cores are mostly used in high frequency applications. Iron cores are made of silicon sheet material and are suitable only for low frequency inductance lines and LV transformer, and are normally used in low frequency and voice frequency applications.
A structure similar to that shown in FIG. 1-2 is a circuit with the switch drive function separated from the main power transformer. As described above, the self-oscillating frequency and drive function of the circuit will be realized by the magnetic saturated transformer B1, therefore, the main power transformer B2 can work in a non-saturated state. Although magnetic saturation occurs at B1, the magnetic saturation consumes very small amount of energy because of the comparatively small volume of B1, and under the identical conditions, the overall efficiency of the Jensen circuit is slightly high.
Presently in the field of switching power, self-excitation push-pull converters are used extensively, because in a small signal model, the only AC small signal input impedance is a positive power, while the input impedance of AC small signals of other switching power is negative, so it is extremely inconvenient to use a shunt switching power with negative input impedance of AC small signals, but the input impedance of AC small signals of self-excitation push-pull converters is positive, very convenient in shunt and cascade applications. But the above-mentioned self-excitation push-pull converters have the following disadvantages:
1. the Converter Efficiency is Low with Light Loads
As the Royer circuit performs push-pull oscillation by using the saturation characteristics of the magnetic core, its static current cannot be too low, and Table 1 shows the measured parameters of the Royer circuit. If a circuit as shown in FIG. 1-1 is used to make a converter with input DC at 5V, output DC at 5V and output current of 200 mA, i.e. with an output power of 1 W. Downstream the transformer, the output is in the circuit structure as shown in FIG. 4, which is a generally known full-wave rectifying circuit, both D41 and D42 are RB160 Schottky diodes. The main parameters of the circuit are: the capacitor C is 1 uF, resistor R1 is 1 KΩ, capacitor C1 is 0.047 uF, and TR1 and TR2 are switching transistors with magnification factor of about 200, with its maximum collector working current being 1 A. The primary side coils NP1 and NP2 have respectively 20 turns, feedback coils NB1 and NB2 respectively 3 turns, secondary side coils NS1 and NS2 respectively 23 turns, and the magnetic core is a common ferrite loop magnetic core with an outer diameter of 5 mm and sectional area of 1.5 mm2, with the common name magnetic loop, and its 3D profile view is as shown in FIG. 5.
In actual measurement, the circuit has a no-load working current of 18 mA, its working frequency is 97.3 KHz, close to 100 KHz, for the conversion efficiency test, the circuit as shown in FIG. 6 was used, VI voltmeter head is working voltage Vin, i.e. the input voltage. A1 ammeter head is input current Iin, i.e. the working current; V2 voltmeter head is output voltage Vout, and A2 ammeter head the output current lout; so the conversion efficiency can be calculated using formula (2). The conversion efficiency of the circuit is:
                    η        =                                            Vout              ×              Iout                                      Vin              ×              Iin                                ×          100          ⁢          %                                    Formula        ⁢                                  ⁢                  (          2          )                    where: Vin is working voltage, i.e. the input voltage, Iin is input current; Vout is output voltage, and lout is output current. In the test, the wiring method as shown in FIG. 6 is used, with RL as the variable load, to effectively reduce the measuring error. Both ammeter and voltmeter are model MY65 4-1/2 digital universal meters of MASTECH® Brand set at steps 200 mA and 20V or 200V, and four and more universal meters were used at the same time.
When the model MY65 4-1/2 digital universal meter is used to measure voltage, the internal resistance is 10 MΩ, and is 200 mA at the 200 mA current step. When the current exceeds 200 mA, two ammeters are used and set at 200 mA to measure it in parallel, and the sum of the current readings of the two meters is the measured value. Measurement using ammeters connected in parallel is a mature existing technology in electronic engineering.
When the circuit as shown in FIG. 1-1 is used and the above-mentioned parameters are set, with the output current at 5% of the full load of 200 mA, or 10 mA, the measured parameters are as shown in Table 1 below.
TABLE 1InputEfficiencycurrentInput voltageOutput currentOutput voltage(CalculatedIinVinIoutVoutvalue)28.4 mA5.060 V9.96 mA5.487 V38.03%
It can be seen from the table above that, when the output is 5% of full load, the efficiency is only 38.03%, which is highly representative in the low power module power sources presently used in the industry. FIG. 2-2 is the measured ripple waveform diagram by overlapping the output DC voltage of the above-mentioned transformer. The test condition is a full load output of 200 mA at a peak value of 135.7.
With the Jensen circuit as shown in FIG. 1-2, although a small transformer B1 is used to realize magnetic saturation while the main power transformer B2 works in a non-saturated state, in an attempt to increase the efficiency, in fact, the use of two transformers results in an additional element to produce loss, and the design of the small transformer B1 should take into account the output power of the whole circuit, after careful commissioning, at 5V output, the no-load current of the Jensen circuit outputting 5V/200 mA is reduced to 16 mA, and when the output is only 5% of full load, the efficiency increases a little from that of the above-mentioned Royer circuit, to 40.91%.
2. It is not Possible to Further Increase the Efficiency at Rated Load
With the self-excitation push-pull converter, take the Royer circuit as an example, as each push-pull operation of the circuit is realized when approaching or at magnetic saturation of the magnetic core, and the energy consumed by magnetic saturation is lost in the form of heat, therefore, to increase the conversion efficiency of the circuit, the working frequency of the converter should be reduced, it can be seen from formula (1) that, with the input voltage remaining unchanged, it can be achieved only by increasing the value of the denominator in formula (1), i.e. increasing the magnetic induction intensity Bw, or increasing the number of coil turns N, or increasing the effective sectional area S of the magnetic core. In converter products today, magnetic cores with extremely high working magnetic induction intensity Bw have been selected, the number of coil turns N increased, resulting in increased copper loss; the increase of the effective sectional area S of magnetic core also increases the loss each time when it approaches or enters the magnetic core magnetic saturation, thus reducing instead of increasing the conversion efficiency of the converter. Therefore, in the design of a self-excitation push-pull converter, it is quite difficult to select between these parameters.
To increase the conversion efficiency of the Jensen circuit, for the similar reasons, if the effective sectional area S of the magnetic core of the small transformer B1 is too small, the pushing power will be insufficient, the switching transistor cannot become well saturated and conducting, resulting in increased voltage drop loss and reduced converter conversion efficiency. When the effective sectional area S of the magnetic core of the small transformer B1 is taken too high, the self-loss will also be high. Of course, the problem insufficient pushing power can be solved by increasing the number of turns N of the coil, but it also results in the following process problem: with increased number of turns N, as the small transformer B1 must work under magnetic saturation state and no air gap can be made, it will make the winding highly difficult.
3. when the Input Voltage is High, there are Many Turns on Transformer B1, Making the Processing Quite Difficult.
In a self-excitation push-pull converter, take the Royer circuit as an example, it can be seen from formula (1) that, when the input voltage increases, if the working frequency of the self-excitation push-pull converter remains unchanged, the corresponding parameter of the denominator in formula (1) should be increased, and for industrial class small module power sources of the same series and same power, magnetic cores of the identical size are often used. In this case, the problem can be solved only by increasing the number of coil turns N, for example, with the circuit parameters shown in FIG. 1-1, if a product with input of 24V is made, the number of turns in the primary side coils NP1 and NP2 should be increased from 20 each for 5V to 96 each, as the transformer B1 in FIG. 1-1 must work in a status allowing magnetic saturation, no air gap is permissible, it is quite difficult to wind the coil, at present, it is quite difficult to wind so many turns of enameled wire on a small magnetic ring with a diameter less than 10 mm, either with a machine or manually. When a machine is used, when the first layer is finished, it is quite difficult to wind the second layer on the first one, as it will break the wire sequence of the first layer, and the winding will be made worse and worse. In manual winding, it is quite difficult to avoid one or two turns more or less as the number of turns must be memorized entirely by workers, if the number of turns is different, deviation will occur in the output voltage, and in a serious case, the original function cannot be realized when the transformer is installed.
If the effective sectional area S of the magnetic core is doubled, the number of turns can be reduced to 48, but in this case, as the effective sectional area S of the magnetic core of transformer B1 is doubled, at the same frequency, the self loss will also double, so the converter conversion efficiency will be reduced.
Therefore, in the industrial field and market today, it is difficult to find self-excitation push-pull converter modules with working voltage of 48V and over, and also for this reason, the efficiency has to be reduced for less number of turns.
4. it is Difficult to Increase the Working Frequency
As the self-excitation push-pull converter circuit performs each push-pull operation by the magnetic core approaching or reaching magnetic saturation, therefore, when the working frequency increases, its loss will increase and conversion efficiency will reduce.
To the Jensen circuit, for the similar reason, the effective sectional area S of the magnetic core of the small transformer B1 become smaller, for a 24V input voltage, it often requires to have 60 turns on the primary side, as there is only one primary side coil, it can be wound with two wires in parallel for only 30 turns, and then they can be connected in parallel as 60 turns, but the small transformer B1 has a smaller diameter, it is quite difficult to wind it either with a machine or manually. For a 48V input voltage, it is almost impossible to make a small transformer B1. Similarly, if the effective sectional area S of the magnetic core is doubled, there can be less turns, but at the same working frequency, the self-loss will double and the conversion efficiency of the converter will reduce.
5. Current Spikes Appear when the Circuit Conducts or is Cut Off, Therefore the Converter Efficiency is Low
In the original book Design of Switching Power published by the Electronic Industry Press, it is also mentioned in the last line on p. 168 that, current spikes appear when the circuit conducts or is cut off, therefore the converter efficiency is reduced to 50.6%. The ISBN No. of the book is 7-121-01755-5. In fact, as most self-excitation push-pull converter for industrial use at present use magnetic core solid surface transformer, for example the conversion efficiency of the 5V/5V product is normally about 78%, but even so, this efficiency is fairly low in the times that energy conservation is advocated.
At present, for self-excitation push-pull converters mentioned in the documents with the file No. 201010604284.7, 201110200894.5, 201110242377.4, 201110247645.1 and 201110272261.5, there is no suitable technology to overcome the afore-said disadvantages of the self-excitation push-pull converters, and these disadvantages are summarized as follows:
1) The converter efficiency is low with light loads;
2) It is not possible to further increase the efficiency at rated load;
3) When the input voltage is high, there are many turns on transformer B1, making the processing quite difficult;
4) It is difficult to increase the working frequency;
5) Current spikes appear when the circuit conducts or is cut off, and therefore the converter efficiency is low.